AI writes code in seconds, but delivery still takes days
AI pushed throughput up 59%, yet median delivery got worse. The bottleneck moved to validation. How replaying production traffic in CI closes the gap.
Posts by Ken Ahrens
Co-founder and CEO of Speedscale, passionate about performance engineering. • 9 posts published
Which Bugs AI Agents Fix Better With Traffic
A second run of our AI bug-fixing benchmark shows where captured traffic lifts agents toward 90%, why service maps barely help, and which bugs still fail.
Build WireMock mappings fast from real traffic
Skip hand-writing WireMock stubs. Speedscale records your real request and response traffic and exports it straight to WireMock mappings.
Fixing 403 auth errors when you replay traffic
Replay an authenticated flow and the protected calls fail with 403. Here is how proxymock recommendations fix the expired bearer token in one click.
Capture once, test forever
Production traffic is the most complete record of what your system does, and most teams throw it away. One capture powers reproduce, validate, and sandbox.
We won't train on your data is not a security architecture
MSA clauses and contractual guarantees aren't an architecture. If your production traffic leaves your cloud, you're trusting a policy, not a system.
Logs told me something broke. Traffic showed me what.
Capture production traffic and store it in your own Elasticsearch with Speedscale BYOC. Pull it locally with es-gather.py and reproduce bugs with proxymock.
Your AI agent is fixing the wrong service
I tested 100 bugs across 240 microservices the model has never seen. Alert only: 51% pass rate, wrong service 34% of the time. Traffic captures: 77%.
WireMock alternatives: pick the one that fits your problem
A 2026 take on WireMock alternatives. Keep WireMock for what it's good at; add Microcks (now CNCF Incubating) or proxymock for the gaps. The honest map.
Try Speedscale free
Choose the desktop proxymock or the hosted cloud trial to get started.